IMF warns global financial system unprepared for AI cybersecurity risks
The IMF warned that the global financial system is unprepared for cybersecurity risks from artificial intelligence, with risks growing exponentially. Anthropic delayed release of its AI model Mythos due to safety concerns about its ability to detect novel computer vulnerabilities at unprecedented scale.
What happened
On 12 April 2026, IMF Managing Director Kristalina Georgieva stated that the international monetary system lacks collective capacity to protect against large-scale AI-driven cybersecurity risks, and that such risks are growing exponentially.
Anthropic delayed the release of its AI model Mythos, which was described as 'very dangerous' in its current state and has the capacity to detect novel computer failures at unprecedented scale and speed. Anthropic requested a consortium of technology companies including Google, Nvidia, Apple and Microsoft to test the model to identify and correct vulnerabilities before potential launch.
Reported impact
Georgieva stated that 'time is not in our favour' regarding protection of the international monetary system against these risks.
Index analysis
Cybersecurity specialists interpreted the warnings about Mythos as the realization of predictions about the rapid rise of AI.
Governance controls
Anthropic requested testing by a consortium of major technology companies to address identified vulnerabilities before launch.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.